@pmatilai commented on this pull request.
> @@ -132,6 +132,8 @@
%_keyringpath %{_dbpath}/pubkeys/
+%specpartsdir %{_builddir}/%{buildsubdir}-SPECPARTS
In practise yes, because the debuginfo machinery depends on it.
--
Reply to this email directly or view it on GitHub:
https://g
@pmatilai commented on this pull request.
> @@ -132,6 +132,8 @@
%_keyringpath %{_dbpath}/pubkeys/
+%specpartsdir %{_builddir}/%{buildsubdir}-SPECPARTS
Oh and just realized that moving the specparts dir out of buildsubdir makes it
kinda unnecessary to make it user overridable, but
@kloczek if you don't know then perhaps you should find out before making such
statements.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2481#issuecomment-1582196320
You are receiving this because you are subscribed to this thread.
Mes
Merged #2481 into master.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2481#event-9469806058
You are receiving this because you are subscribed to this thread.
Message ID:
___
Rpm-maint maili
Of course, moving it out of the main build directory has the same caveat as
renaming to a dotfile: the magic becomes largely invisible to the packager.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2534#issuecomment-1582200853
You are r
Yes, the underscores in macros are a bit of a mess. I modeled %specpartsdir on
%buildsubdir which also comes without an underscore. I think the idea is that
macros to be used directly don't have an underscore. Stuff to be configured
have one underscore and stuff not to be touched have two.
--
The first commits are just refactoring preliminaries, the beef is in the
second-last commit which adds a new opt-in generator mode where all the files
are processed in one go per extracted dependency type instead of launching a
new generator for every file separately.
This is a several orders o
@pmatilai pushed 7 commits.
34a1f71ad7261bd1620ae71009208d283cdbb7c9 Refactor file walk loop into
rpmfcHelper()
79f03fede68bf06c8604aeaf1dcf390558b2bc4c Pass the entire attribute to
applyAttr() for more flexibility
e6f9d37c02d28d9bc1e6f5f3f048d8ef75be2a58 Split weeding out excludes to a
sepa
That parallelization argument is getting really old you know. You've repeated
it many, many times and I've repeatedly told you it's not that easy, and it
would not make it any less horribly inefficient, even if it managed to consume
all 256 cpus at once.
I just submitted the multifile-dependenc
The Python generator has always supported multifile, since it was originally
written for that case way back when. This will be a sweet speed-up!
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2537#issuecomment-1582302318
You are receivin
For an added bonus, in the multifile mode a generator can internally
parallelize operation if it so wishes.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2537#issuecomment-1582302733
You are receiving this because you are subscribed to
> The Python generator has always supported multifile, since it was originally
> written for that case way back when. This will be a sweet speed-up!
Yeah, pratically all generators were written to support multiple files as input
because that's how the external dependency generator worked. The on
> @kloczek if you don't know then perhaps you should find out before making
> such statements.
As I wrote %set_build_flags is using $CFLAGS, $CXXFLAGS, $LDFLAGS, $NM, $CC,
$CXX and so on because those env variables comes with meson, GNU, autotools,
cmake and even plain make.
If some project is
Wouldn't be easier to detect the `;` at the beginning of line instead of the
explicit `%__elf_protocol multifile`?
BTW it would help if the commit messages were extracted into some documentation.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rp
Just FTR, the
[Ruby](https://src.fedoraproject.org/rpms/ruby/blob/rawhide/f/rubygems.req)
[dependency](https://src.fedoraproject.org/rpms/ruby/blob/rawhide/f/rubygems.prov)
[generators](https://src.fedoraproject.org/rpms/ruby/blob/rawhide/f/rubygems.con)
are almost ready to use this. However,
> Wouldn't be easier to detect the `;` at the beginning of line instead of the
> explicit `%__elf_protocol multifile`?
Well no, that's a day too late. Rpm needs to calculate the files sent to the
generator before the fact, and for that a signal in the response is useless.
> BTW it would help if
@pmatilai pushed 1 commit.
7e9a19baccfdb588f2db95fe8991e9949b0adfce fixup! Support optional multifile
protocol in dependency generation
--
View it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2537/files/3ce09fa9a8d0ba7cf26709bbc182c7e6dadb2090..7e9a19baccfdb588f2db95fe8991e99
BTW will old RPM cope with the `%__NAME_protocol` in .attr file?
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2537#issuecomment-1582519465
You are receiving this because you are subscribed to this thread.
Message ID: __
Old rpm will just silently ignore the protocol setting.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2537#issuecomment-1582547519
You are receiving this because you are subscribed to this thread.
Message ID: ___
@rhabacker pushed 2 commits.
707ea38dcd1fcdb75377c046dac42ae7047a9e32 Add macro '%-x**' containing all
occurrences of the flag '-x' or '-x '.
a1476b8610e42db5c079bd268bf975fd4a1eb6ab Add test case for macro '%-x**'
--
View it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2449
20 matches
Mail list logo
